Dear all,

Referring to the description of reaction in 319.5 of B31.3-2014, it has been changed to say that "the reactions shall be calculated using the modulus of elasticity AT THE TEMPERATURE OF THE CONDITION, Em (Ea may be used instead of Em when it provides a more conservative result)".

I think the code now requires that the default calculation of reaction shall be based on Em, not Ea, because usign Ea will not "always" give conservative result.
So I am wondering whether I should use EH1, EH2,... for analysis, to incorporate code requirement correctly.

Please share your opinion for better understanding.

Thank you.

Expansion stress calculations should always use the reference modulus of elasticity - Ea. Markl's work cycled components to failure using a displacement at ambient conditions so our analysis (based on Markl's "SIF" approach) must not use a modulus that would reduce the load associated with the expansion stress range.
Calculating system reactions is not directly related to the Markl work and does not have the same restriction - you can use the modulus associated with the analysis temperature.
The "Reactions" paragraph in B31.3 could use an update to take us beyond those simple slide rule estimates.

to summarize, the stress calculations trump the reaction calculations - use Ea for stress but you may also take a look at Em for reactions.

Thank you Mr. Diehl.

As you pointed out, B31.3-2014 still requires the stress range to be calculated using Ea, as per paragraph 319.4.4 (a).

Going back to the paragraph 319.5, I still doubt that ;

if the purpose of code change was just to allow code users to have options to use Em, instead of Ea, then the code should have said that "Ea shall be used, but Em may be used". But, the code changed to say that "Em shall be used, but Ea may be used".

In most cases using Ea will give us more conservative results in calculating reactions than Em will, but the difference in calculated reactions between using Ea and using Em, sometimes seems to be excessive to be considered as just "conservatism".

I think to be "conservative" in designing reactions, it would be better to have a safety factor to calculated loads, not to use higher elastic modulus.

Thank you.

If you want to do this you will have to:

1) Duplicate the load case in the Load Case Editor.
2) For this duplicate load case, change the modulus to EHn as appropriate.
3) Make sure you don't use any of the results from this load case - except for the restraint loads. Do not use this load case in further "load case combinations".
4) I suggest you use the "load case name" option to label this duplicate case appropriately to avoid confusion in the future.